Sunset Mesa Software

Make AI text sound like you wrote it

Rewire Text's Humanize transform rewrites AI-generated text so it reads like a person wrote it: the em-dashes go, the stock phrases go, the metronome rhythm varies, and the meaning stays put. It works in any app (select the text, press the hotkey, run Humanize) and it can run entirely on your own machine through a local model. No website with a paste box, no per-word credits. And when 'sounds like a person' is not enough, a saved custom transform can aim at the real target: sounds like you.

Why AI text reads as AI text

The tells are consistent enough that most readers now spot them: an em-dash in every second sentence, "it's not just X, it's Y", triads everywhere ("faster, simpler, and more reliable"), and paragraphs whose sentences all land with the same beat. None of that is wrong, exactly. It just is not how you write, and people notice.

The fix is not asking a chatbot to "make it sound human" and pasting the result back and forth. Humanize is a single transform in the palette, applied to the text right where it is: an email draft, a doc, a ticket comment, a slide.

Humanize AI transform. Output is representative; results vary by model.

Before

In today’s fast-paced digital landscape, clear communication isn’t just important—it’s essential. Our new onboarding flow doesn’t merely simplify setup—it transforms the entire first-run experience.

After

Clear communication matters, and onboarding is where new users feel it first. The new flow cuts setup down to a few steps and makes the first run a lot less confusing.
Humanize AI transform. Output is representative; results vary by model.

Before

The update ships Friday—assuming QA signs off—and includes the new export flow, improved error handling, and enhanced performance.

After

The update ships Friday, assuming QA signs off. It includes the new export flow, better error handling, and some performance work.

One step further: your voice, not just a human one

Humanize is deliberately generic. It removes the machine tells and leaves clean, neutral prose, which is the right target for a ticket comment or a status doc. But "sounds like a person" and "sounds like you" are different goals. Your writing has habits: how long your sentences run, the words you never use, how quickly you get to the point, how much hedging you tolerate.

For that, save a custom transform. You write the instruction once, the way you would brief an editor, give it a name, and it becomes a transform in the palette like any built-in one. For example:

Custom: Rewrite in my voice The saved instruction. Yours will look different; that is the point.
Rewrite this in my voice. Plain words, short sentences, contractions. Get to the point in the first sentence, no warm-up phrases. When I hedge, I say "probably" and move on. Never use "leverage", "streamline", or "excited to share". Keep technical terms and numbers exactly as written. One dry aside is fine if it earns its place.

Run on the same text as the first example above, the difference is easy to see:

Custom: Rewrite in my voice AI transform. Output is representative; results vary by model.

Before

In today’s fast-paced digital landscape, clear communication isn’t just important—it’s essential. Our new onboarding flow doesn’t merely simplify setup—it transforms the entire first-run experience.

After

The new onboarding flow is live. Setup is down to a few steps, and new users shouldn’t get lost on their first run anymore.

Custom transforms live in Settings, under Transforms, where you can edit the instruction as you notice more of your own habits, and choose how capable a model each one runs on. Save as many as you like: one for your email voice, one for commit messages, one for the way your team writes release notes.

On your desktop, optionally offline

The web is full of "AI humanizer" sites with paste boxes, word limits, and their own metering. Rewire Text is a desktop utility: the transform runs from a hotkey in whatever app you are in, with your own API key (Gemini, Claude, OpenAI, Grok, OpenRouter, or any OpenAI-compatible endpoint) or with a local model through Ollama, LM Studio, or llama.cpp, in which case the text never leaves your machine. Confidential drafts stay confidential.

And the honest framing, stated plainly: use it on text you are entitled to present as yours, to make it read the way you actually write. If the goal is fooling an AI detector, this is the wrong tool, and mostly the wrong goal.

Transforms for this workflow

  • Humanize
  • Rewrite
  • Make Casual
  • Improve Clarity
  • Custom Instruction
  • Custom: Rewrite in my voice

Part of 50+ built-in transforms. See the full list

Questions

Will this bypass AI detectors?

That is not what it is for, and no tool can honestly promise it. Humanize fixes style: the em-dashes, the stock phrases, the uniform rhythm that make text read as machine-written. Judge the result by whether it sounds like you, not by a detector score.

Does Humanize change what the text says?

No. It preserves the meaning and the language and returns only the rewritten text. It removes the tells, not the content.

Can it run offline, without sending my text to a cloud service?

Yes. Humanize is an AI transform, so point Rewire Text at a local model through Ollama, LM Studio, or llama.cpp and the whole thing runs on your machine. With a cloud key instead, the text goes only to the provider you chose.

What exactly are the "AI tells" it removes?

Em-dashes sprinkled through every paragraph, constructions like "it’s not just X, it’s Y", the rule-of-three cadence on every line, stock openers like "in today’s fast-paced world", and sentence rhythms that never vary. Humanize rewrites around them.

Is this one click, or do I write a prompt each time?

One click. Humanize is a fixed transform in the palette: select text, press the hotkey, type "hum", press Enter. A one-off personal touch ("make it drier", "match my usual sign-off") can be a free-form Custom Instruction. And guidance you keep repeating belongs in a saved custom transform, which makes it one click too.

Can it match my personal style, not just sound generically human?

Yes, through a custom transform. There is no training and no watching you write: you describe your style once (sentence length, words you avoid, how direct you are) and save it as a named transform. A paragraph of honest guidance gets most of the way there, and you can sharpen the instruction whenever you notice another habit.

Try it on your own text

Download Rewire Text and use everything free for 7 days. Buy it once for $29 and it runs on any two of your devices, macOS or Windows.

Learn more about Rewire Text